[PATCH 4/4] gdb/fortran: Fix sizeof intrinsic for Fortran

[Kempke, Nils-Christian]

> -----Original Message-----
> From: Simon Marchi <simark@simark.ca>
> Sent: Monday, September 26, 2022 7:07 PM
> To: Kempke, Nils-Christian <nils-christian.kempke@intel.com>; gdb-
> patches@sourceware.org
> Cc: tom@tromey.com
> Subject: Re: [PATCH 4/4] gdb/fortran: Fix sizeof intrinsic for Fortran
> 
> 
> 
> On 2022-09-20 03:26, Nils-Christian Kempke via Gdb-patches wrote:
> > The sizeof operator in Fortran behaves differently from e.g. C/Cpp in
> > that it can be applied to pointers.  We thus dereference pointers before
> > evaluating their size.  A test has been added for the Fortran sizeof
> > operator.
> > ---
> >  gdb/eval.c                           |   3 +
> >  gdb/testsuite/gdb.fortran/sizeof.exp | 110
> +++++++++++++++++++++++++++
> >  gdb/testsuite/gdb.fortran/sizeof.f90 | 108
> ++++++++++++++++++++++++++
> >  3 files changed, 221 insertions(+)
> >  create mode 100644 gdb/testsuite/gdb.fortran/sizeof.exp
> >  create mode 100644 gdb/testsuite/gdb.fortran/sizeof.f90
> >
> > diff --git a/gdb/eval.c b/gdb/eval.c
> > index ce1d883aa86..bb6b757d452 100644
> > --- a/gdb/eval.c
> > +++ b/gdb/eval.c
> > @@ -2730,6 +2730,9 @@ evaluate_subexp_for_sizeof_base (struct
> expression *exp, struct type *type)
> >    if (exp->language_defn->la_language == language_cplus
> >        && (TYPE_IS_REFERENCE (type)))
> >      type = check_typedef (TYPE_TARGET_TYPE (type));
> > +  else if (exp->language_defn->la_language == language_fortran
> > +	   && type->code () == TYPE_CODE_PTR)
> > +    type = check_typedef (TYPE_TARGET_TYPE (type));
> 
> Do you need yo handle TYPE_CODE_REF for fortran too?
> 
> >    return value_from_longest (size_type, (LONGEST) TYPE_LENGTH (type));
> >  }
> >
> > diff --git a/gdb/testsuite/gdb.fortran/sizeof.exp
> b/gdb/testsuite/gdb.fortran/sizeof.exp
> > new file mode 100644
> > index 00000000000..f353e8c4dd9
> > --- /dev/null
> > +++ b/gdb/testsuite/gdb.fortran/sizeof.exp
> > @@ -0,0 +1,110 @@
> > +# Copyright 2022 Free Software Foundation, Inc.
> > +
> > +# This program is free software; you can redistribute it and/or modify
> > +# it under the terms of the GNU General Public License as published by
> > +# the Free Software Foundation; either version 3 of the License, or
> > +# (at your option) any later version.
> > +#
> > +# This program is distributed in the hope that it will be useful,
> > +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> > +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> > +# GNU General Public License for more details.
> > +#
> > +# You should have received a copy of the GNU General Public License
> > +# along with this program.  If not, see <http://www.gnu.org/licenses/> .
> > +
> > +# Testing GDB's implementation of SIZE keyword.
> > +
> > +if {[skip_fortran_tests]} { return -1 }
> > +
> > +standard_testfile ".f90"
> > +load_lib fortran.exp
> > +
> > +if {[prepare_for_testing ${testfile}.exp ${testfile} ${srcfile} \
> > +	 {debug f90}]} {
> > +    return -1
> > +}
> > +
> > +if ![fortran_runto_main] {
> > +    return -1
> > +}
> > +
> > +gdb_breakpoint [gdb_get_line_number "Test breakpoint"]
> > +gdb_breakpoint [gdb_get_line_number "Past unassigned pointers"]
> > +gdb_breakpoint [gdb_get_line_number "Final breakpoint"]
> > +
> > +set done_unassigned 0
> > +set found_final_breakpoint 0
> > +set test_count 0
> > +while { $test_count < 200 } {
> > +    with_test_prefix "test $test_count" {
> > +	incr test_count
> > +
> > +	gdb_test_multiple "continue" "continue" {
> > +	    -re -wrap "! Test breakpoint" {
> > +		# We can run a test from here.
> > +	    }
> > +	    -re -wrap "! Past unassigned pointers" {
> > +		# Done with testing unassigned pointers.
> > +		set done_unassigned 1
> > +		continue
> > +	    }
> > +	    -re -wrap "! Final breakpoint" {
> > +		# We're done with the tests.
> > +		set found_final_breakpoint 1
> > +	    }
> > +	}
> > +
> > +	if ($found_final_breakpoint) {
> > +	    break
> > +	}
> > +
> > +	# First grab the expected answer.
> > +	set answer [get_valueof "" "answer" "**unknown**"]
> > +
> > +	# Now move up a frame and figure out a command for us to run
> > +	# as a test.
> > +	set command ""
> > +	gdb_test_multiple "up" "up" {
> > +	    -re -wrap "\r\n\[0-9\]+\[ \t\]+call test_sizeof \\((\[^\r\n\]+)\\)" {
> > +		set command $expect_out(1,string)
> > +	    }
> > +	}
> > +
> > +	gdb_assert { ![string equal $command ""] } "found a command to
> run"
> > +
> > +	set is_pointer_to_array [string match "sizeof (*a_p)*" $command]
> > +
> > +	if {$done_unassigned || !$is_pointer_to_array} {
> > +	    gdb_test "p $command" " = $answer"
> > +	} else {
> > +	    # Gfortran, ifx and ifort have slightly differnt behavior for
> > +	    # unassigned pointers to arrays.  While ifx and ifort will print 0
> > +	    # as the sizeof result, gfortran will print the size of the base
> > +	    # type of the pointer/array.  Since the default behavior in GDB was
> > +	    # to print 0 we keep this and make an exception for gfortran here.
> > +	    gdb_test_multiple "p $command" "p $command" {
> > +		-re -wrap " = $answer" {
> > +		    pass $gdb_test_name
> > +		}
> > +		-re -wrap " = 0" {
> > +		    pass $gdb_test_name
> > +		}
> > +	    }
> > +	}
> > +    }
> > +}
> 
> Can you add some comments to explain what is going on here with the big
> loop?  It's a bit of an unusual structure, so it would help to explain
> it.
> 
> Simon

Sure, will be done in V2!

Thanks for the review!
Nils

Intel Deutschland GmbH
Registered Address: Am Campeon 10, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de <http://www.intel.de>
Managing Directors: Christin Eisenschmid, Sharon Heck, Tiffany Doon Silva  
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928